Congratulations for your article is very interesting and well treated. The paper is an important example of using alternatives to solid precipitation measurements in operational networks taking into account cost-value decisions. It is well suited to the audience of the journal and worth being published. 

I like to remark the good experimental design and the strong methodology with 3 different sites for intercomparison. In my opinion, this is very important because it shows that replication in other sites is possible.  The results are clear and the quality control of data is consistent with other similar studies 

However, some details and explanations are required in order to reproduce the LPDF in other sites, either for other intercomparisons or just for operational measurements.

I would suggest to include a more detailed description of the LPDF including a complete diagram (i.e cross section) and some details how the slats are installed within the chain link fence panels (before/after), how to avoid displacement of LPDF under strong winds, approximate dimensions of concrete footing and  how to anchor and elevate the LPDF etc 

Another important point for discussion is to analyze if given the objective of 25% of porosity the design can/could be slightly different (i.e panel dimensions, slats width, etc) obtaining similar results

Another important point is a more detailed discussion about durability of LPDF compared with SDFIR and the approximate price different on material and time for installation

Also some minor comments are provided below that I would ask the authors to consider before the paper can be accepted for publication.

1)    Remove DFIR on the caption for figure 6 y 8

2)    Percentage of cases Ugh> 9 m/s

In this manuscript Kochendorfer et al. present a new type of precipitation gauge wind shield to mitigate undercatch of solid precipitation in windy conditions. The study describes the Low Porosity Double Fence (LPDF) which is smaller, more durable and easier to install and maintain in remote locations then the current reference-quality wind shields (such as the DFIR and the SDFIR) ((small) Double Fence Intercomparison Reference). I believe the work therefore presents a substantial new method and would recommend the publication of this manuscript after some minor revisions.

Below some more detailed remarks:

Page 1, lines 25-25: “This new wind shield is much smaller and easier to install and maintain” – is it also cheaper? In many organisations (and increasingly so) budget constraints can be an important factor.

Page 2, line 64: perhaps the authors could briefly define/explain the meaning of porosity in the context of windshields, and how it is estimated/calculated.

Page 3, line 95-97: “the design of the LPDF also allows it to be raised much more easily than the DFIR or SDFIR” – here and in general, it would be nice to include a technical drawing of the LPDF. This would also improve reproducibility of the work.

Page 4, site selection. Regarding the site selection procedure, how representative are the chosen sites for USCRN sites overall? I’m also missing a map with the site locations as well as wind statistics for the Marshall site in Figure 1. (consider having a,b and c panels in Figures 1 and 2 to be referring to the same sites).

Page 5, lines 142-155: In this paragraph the authors claim hourly and daily precipitation measurements are subject to random errors and therefore not appropriate for intercomparison studies. This seems in contradiction with paragraph 3.4 (and figures 10 and 11) where hourly catch efficiency measurements are shown. It also seems at odds with the introduction (page 2, lines 33-35) where several short term consequences of precipitation are mentioned. For avalanche predictions for example, accurate precipitation estimates at the precipitation event level would be more useful. Have the authors considered comparisons at the (long) event scale for example?

Page 5, line 152: why is a comparison of long-term accumulation more demanding? Surely on the contrary, it has the effect of averaging out small errors?

Page 6, line 167: “serviced with oil” might be a confusing term here. Perhaps a brief description in the methods section on the anti-freeze (if used) and oil layer - and why these are used, could be helpful to readers unfamiliar with weighing gauges.

Page 6, line 178: could the authors comment on how prevalent situations were in which the SDFIR accumulated more than 0.25 mm in an hour and LPDF less than 0.25 mm, and vice versa? Was this negligible?

Page 7, line 197-200: again, what percentage of total measurements did this represent? Based on Fig 2, windspeeds above 9 m/s must have been a very rare phenomenon.

Page 7-8 paragraph 3.2 and Figure 7: Single events seem to have most affected the difference between the accumulations of solid and mixed precipitation in Chatham. Have the authors looked into this or could they comment on this?

Minor

Page 3, line 1: In addition .. additional (repetition).

Page 4, line 102: “conterminous” not sure what conterminous is supposed to mean in this context – or why it is important that USCRN sites are conterminous.

Page 4, lines 105-108: “ Wind speeds during the snow days .. USCRN sites (e.g. Fig. 2).” These sentences seem unnecessarily repetitive, please consider revising them.

Page 7, lines 216-217: consider adding causal link between these two sentences for better comprehension, i.e. “precipitation shown in Fig.4. This is because the phase discriminated measurements” etc.

Page 8, lines 231-232: consider reporting SDFIR and LPDF values in the same order throughout the manuscript.